Like direct strikes to mainland china - direct strikes to continental us/ us territories imply a global conflict. Article 5 would be called on, and NATO would be drawn into the conflict. China would struggle against a G7 boycott.
Itβs plausible that EU stays neutral in a narrow war, but a broad war with expansive war goals would lead to a broad response. Chinas past behavior towards nuclear powers occurred before they had indigenous weapons and while they had a nuclear protector.
The only limiter to nuclear escalation in a modern global conflict is willingness and alternatives. When alternatives have been removed and civilian targets are targeted nuclear options will be on the table. In a broad conflict this process could occur within days.
Current hardware delivery aren't coming through during active hostilities. Unlike RU who doesn't have capabilities to deny vast UKR + EU land logistics, PRC can proactively deny air+maritime around TW. Optics of PRC allowing foreign powers to actively support seperatist gov on Chinese territory during active Chinese civil war which includes TW soil and 12nm territorial waters is not something tolerable for domestic PRC politics. Or that TW cannot be supplied during war time without invading legal Chinese soil something PRC sovereignty politics won't allow due to history. That's the significant political consideration, UKR supporters moving supplies through soverign UKR territory aren't legal belligerants, supplyiers of TW that necessarily involves what PRC percieves to be invading Chinese territory are.
>imply a global conflict
Yes as does blockading PRC, that's where the entire US can choke PRC strategy falls apart, it now feasibly escalates to CONUS war. If US blockades PRC (i.e. US initiates legal war), then PRC reciprocating attack on CONUS no longer falls under art5. Of course NATO could still join, but with what hardware, based where? Even G7 boycott weatherable, this isn't 2010s when PRC was relatively export dependent, export GDP to west is now barely 10%, with mutual knock on supply chains. Most of G7+ are MUCH more trade dependant than PRC as well and will hurt more, i.e. even least trade dependant US is going to have to contest with trillions lost in advanced industries because TW leading edge semi gone, and likely crippled domestic US semi because they will still be dependant on east asian semi supply chains that's not going to survive a broad war. Meanwhile most of that hardware is already sanctioned / denied to PRC. This doesn't suggest PRC wants a broader war, but broader war does open up host of strategic reach goals and opportunities that asymmetrically devastates PRC containers.
>before they had indigenous weapons and while they had a nuclear protector
Yes, PRC had no issue tussling with nuclear powers while not being nuclear power herself. And she tussled with her alleged nuclear protector USSR as well. Multiple times + other geopolitical drama, all less important than TW.
> civilian targets are targeted
This excludes TW civilian targets (and likely strategic targets on JP, SKR, PH, AU etc), as old addage goes, is US willing to trade nukes for Los Angeles over Taipei. Proxies can fight to the last man / brick. If US doesn't want to escalate to nuclear with PRC then the point of CONUS retaliatory strikes is to deter US from dirupting PRC civilian homefront, which includes PRC shipping. But really I'd say esclation ladder jumps near nuclear once homeland attacks, even if just military are happening, which many US war planners still think is allowable - plenty of entrenched hubris thinking unilaterally attacking PRC mainland targets would not invite proportional retaliation on CONUS or other belligerants against PRC. Chance of US blockading PRC or hitting mainland PRC increasingly unlikely as PRC fills out conventional CONUS escalation rungs. Intervention calculus changes when there's homeland vunerability, this includes US allies whose critical infra is vunerable to PRC global strike as well. If anything home land vunerability limit nuclear escalation by stopping escalation at critical conventional rungs and limit broadness of response. Alternatively this is also a world where conventional escalation can reach near existential levels of mutually assured destruction. Imagine escalation ladder with rung where US+co refinery goes boom and first world economies simply ceases to be viable. That's what's skewing escalation dominance towards PRC over core PRC concerns over TW, which starts at preventing foreign aid to speed up TW collapse/capitulation.
>they cannot win
PRC can very well win against US regional allies, and in many ways, a broader war is preferable because limited war means US IndoPac security infra remains untouched and others in region free to hedge with US security after TW damocles off the table. The greatest PRC strategic vision is to kick US security out of east asia and for that she needs a broad war where PRC demonstrates US cannot fulfill security commitments by destroying a protectorate so utterly that no one else would consider bandwagoning under US security umbrella after TW. Even in case of draw/loss, having opportunity to knock out extensive US IndoPac military infra (i.e. PineGap in AU) cripples US threatre capability for potentially decades. Who wins is relative, one side can win more. PRC and US both losing similar amount of blue water hulls = PRC can STILL assert military projection in her backyard with land based air assets, VS US basically gets kicked out of the Asian power projection and perhaps global hegemony game. The loss is not symmetric.
Also PRC doesn't need USN to protect global shipping, that's more Zeihan wank that hasn't been relevant for decade+. What PRC has been doing is freeloading off US, but it's more than capable to assuming own protection duties.
I've wrote elsewhere about how seriously PRC historically takes nuclear escalation with respect to security interests - it doesn't - having fought with US, USSR, IND, threatened UK, proxy FR over things much less important in TW. US carrier groups are unlikely to threaten mainland with PRC a2ad which again extends to Guam, well past even CSG sprint + stand off extended range. And even lul recent balloon drama aside, PRC rocket force can reciprocate hitting CONUS infra conventionally. If the US takes war to PRC mainland, expect critical infra to be gone on CONUS. It will take a few years for mutual vulnerability to filter way through western analysis, but all the US wank expeditionary warfare is contingent on it being unilateral, with homeland remaining un-disrupted. That's not true with PRC anymore, with PRC ability to strike CONUS only increasing with time, which will greatly circumscribe what the US will do in the TW scenario. And with precise rapid strikes being developed (PRC parlance for prompt global strike), we're talking about every US capital asset being a port visit away from being scrap. Including upcoming B21s that still needs hanger dramaqueen time. West keeps focusing on US deterrence efforts around TW without realizing PRC is also simultaneously building up intervention deterrence. Except PRC deterrence development is much more significant - US could always hit targets in mainland, that calculation was baked it. It's when PRC can hit CONUS that real deterrence math changes.